The standard module and its left and right Hecke operators
Fix a Coxeter system \((W,S)\) and a commutative ring \(C\). The standard module is \[ \mathcal H_C=\bigoplus_{w\in W}Ce_w. \] A vector is a function \(f:W\to C\) with finite support, written \(\sum_w f(w)e_w\). Its coefficient at \(w\) is \(f(w)\); \(e_w\) has coefficient one at \(w\) and zero elsewhere. This is the source's finitely supported function type.
An operator is a \(C\)-linear map from this module to itself. Specifying its values on the \(e_w\) defines it on a general vector by \[ L\left(\sum_w f(w)e_w\right)=\sum_w f(w)L(e_w). \] The sum is finite even when \(W\) is infinite.
The normalized convention
For \(a\in C\), use the basis notation \(H_w=e_w\). The left operator is \[ L_s(a)H_w= \begin{cases} H_{sw},&\ell(sw)>\ell(w),\\ H_{sw}+aH_w,&\ell(sw)<\ell(w). \end{cases} \] The right operator \(D_t(a)\) uses \(wt\) in place of \(sw\). Simple multiplication changes length by exactly one, so these two cases exhaust the possibilities.
The symbols \(L_s,D_t\) name operators; \(H_w\) names a vector. The expression \(H_sH_w\) abbreviates the left operator value in the regular-action notation.
On a pair \(H_w,H_{sw}\) with \(\ell(sw)=\ell(w)+1\), the matrix is \[ \begin{pmatrix}0&1\\1&a\end{pmatrix}. \] The first column says \(H_w\mapsto H_{sw}\); the second says \(H_{sw}\mapsto H_w+aH_{sw}\).
The unnormalized convention
For \(q\in C\), write \(T_w=e_w\) and use \[ L_s(q)T_w= \begin{cases} T_{sw},&\ell(sw)>\ell(w),\\ qT_{sw}+(q-1)T_w,&\ell(sw)<\ell(w). \end{cases} \] Again \(D_t(q)\) is the right version. On an ascent pair its matrix is \[ \begin{pmatrix}0&q\\1&q-1\end{pmatrix}. \] These operators are defined for every \(q\). Their inverse construction uses an invertible \(q\).
Conversion over Laurent coefficients
Take \(q=v^2\) with \(v\) invertible, and set \(H_w=v^{-\ell(w)}T_w\). The operator \(v^{-1}L_s(q)\) then has normalized parameter \(a=v-v^{-1}\).
In an ascent, \(v^{-1-\ell(w)}T_{sw}=H_{sw}\). In a descent, \(\ell(sw)=\ell(w)-1\), so \[ v^{-1-\ell(w)}\bigl(v^2T_{sw}+(v^2-1)T_w\bigr) =H_{sw}+(v-v^{-1})H_w. \] Thus the two tables describe the two coefficient conventions used in the source. Later formulas specify which parameter and basis are being used.
References
The local source declarations below give the construction and proof.
Lean formalization
OAI.KLInvariance.Hecke.Space: Hecke.lean:24.
OAI.KLInvariance.Hecke.basis: Hecke.lean:26.
OAI.KLInvariance.Hecke.left: Hecke.lean:39.
OAI.KLInvariance.Hecke.right: Hecke.lean:44.
OAI.KLInvariance.HeckeQ.left: Hecke.lean:514.
OAI.KLInvariance.HeckeQ.right: Hecke.lean:521.
Used by
- The quadratic relation for a Hecke generator
- Quadratic identities and inverses of simple Hecke operators
- Word-stalk characters equal positive Hecke-word coefficients
- Left and right simple Hecke operators commute
- The product recurrence for dihedral lower-interval sums
- Lower-interval Hecke sums and their correction vectors
- Positive Hecke words and their coefficient recurrence
- Word operators and the standard invertible action
- Hecke operators, the bar map, and normalized R-polynomials